Amanda-Users

Re: Backup too long with windows client

2005-08-17 13:22:52
Subject: Re: Backup too long with windows client
From: Paul Bijnens <paul.bijnens AT xplanation DOT com>
To: tanguy yoann <tanguy_yoann AT yahoo DOT fr>
Date: Wed, 17 Aug 2005 19:13:27 +0200
tanguy yoann wrote:
Hi,

        I have a problem with a backup:  it puts too much
time.  I do not backup many data but that takes a long
time.  I made this backup before and that lasted only
a few seconds.  I backup a Windows share by smbclient.
 I do not understand the problem. So I send to you

To be honest, I do not understand the problem either.
Try to reword the problem and be more specific in
what you expect and what you get (instead of "a few"
and "a long time").


files of debug to see whether you will be able to
understand what does not function correctly. Thanks a
lot.

...
GNUTAR //Yoann/backup 1 2005:8:18:7:54:1 -1 OPTIONS
|;bsd-auth;compress-best;index;

First candidate:
"compress client best" compresses only a few percentages
more, but takes twice or more time and cpu.
Use "compress client fast" instead.


...
amandad: time 0.000: running service
"/usr/lib/amanda/sendsize"
amandad: time 239.967: sending PREP packet:
----
Amanda 2.4 PREP HANDLE 000-18820608 SEQ 1124436519
OPTIONS features=fffffeff9ffe7f;
----


Second wierd thing above:  why is amanda waiting 239
seconds before sending the first PREP (partial reply)
package?
On one of my slowest clients it takes 0.06 seconds.
This is one of the newer features in 2.4.5, which I'm
not completely familiar with.  If I understand the
code, then the first packet, containing only the
"OPTIONS feature=..." is send immediately after
the sendsize invocation.
If that is correct, than why does fork/exec of sendsize
take 239 seconds?


amandad: time 360.013: sending PREP packet:
...
amandad: time 360.070: pid 6666 finish time Fri Aug 19
09:34:38 2005

And 120 seconds later the "sendsize" is complete.
Is 120 seconds a normal time to be expected here?
(Above you said "a few seconds".)



2)amandad sendbackup

...
amandad: time 0.000: running service
"/usr/lib/amanda/sendbackup"
amandad: time 0.003: sending REP packet:
----
Amanda 2.4 REP HANDLE 000-E8ED0608 SEQ 1124436518
CONNECT DATA 50084 MESG 50085 INDEX 50086
OPTIONS features=fffffeff9ffe7f;
----

amandad: time 0.006: got packet:
----
Amanda 2.4 ACK HANDLE 000-E8ED0608 SEQ 1124436518
----

amandad: time 0.006: pid 6709 finish time Fri Aug 19
09:34:38 2005

Very weird that the "sendbackup" takes 0.003 seconds
while the sendsize took over 120 seconds (or even 360
seconds).




3)Sendsize

...
sendsize[6706]: time 359.946: spawning
/usr/bin/smbclient in pipeline
...
sendsize[6706]: time 360.008: Total number of bytes:
260218
sendsize[6706]: time 360.009: .....
sendsize[6706]: estimate time for //Yoann/backup level
1: 0.063

The real estimate took a fraction of a second, as
expected (smbclient uses the internal "du" function
for estimates).


4)Sendbackup
...
sendbackup-gnutar: time 0.004: pid 6713: /bin/gzip
--best
sendbackup-gnutar: time 239.968: doing level 1 dump
from date: 2005-08-29  7:00:19 GMT
sendbackup-gnutar: time 359.950: backup of
\\Yoann\backup

Notice again the 239 seconds - 360 seconds milestones
before sendbackup actually does something useful.
Wierd, but you know your system better than I do...


--
Paul Bijnens, Xplanation                            Tel  +32 16 397.511
Technologielaan 21 bus 2, B-3001 Leuven, BELGIUM    Fax  +32 16 397.512
http://www.xplanation.com/          email:  Paul.Bijnens AT xplanation DOT com
***********************************************************************
* I think I've got the hang of it now:  exit, ^D, ^C, ^\, ^Z, ^Q, F6, *
* quit,  ZZ, :q, :q!,  M-Z, ^X^C,  logoff, logout, close, bye,  /bye, *
* stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt,  abort,  hangup, *
* PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e,  kill -1 $$,  shutdown, *
* init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... *
* ...  "Are you sure?"  ...   YES   ...   Phew ...   I'm out          *
***********************************************************************



<Prev in Thread] Current Thread [Next in Thread>